A Discussion on Explicit Knowledge

In the papers on knowledge during the last decade, a recurrent theme is the distinction between implicit knowledge ascribed to an agent and explicit knowledge available to that agent. We discuss one aspect of this distinction: even a logically omniscient reasoner's knowledge may be limited by having only a partial view of the system. We study a logic where an agent has explicit knowledge of only those facts which hold in worlds visible to the agent. 1 A distinction During the last decade, there has been extensive research into the notion of ascribing knowledge to reasoning agents, particularly in the contexts of distributed systems, ar-tiicial intelligence and mathematical economics. The recent book FHMV95] provides an excellent summary of dominant themes in this line of research. Following Hintikka's logical treatment, knowledge is studied as a modality in (typically) a propositional modal logic, with the semantics of formulas given in Kripke frames where the accessibility relation is an equivalence. Formally, we have a count-able set of propositions P = fp 0 ; p 1